Skip to Main Content
In this paper, we describe a real-time gradient-based multiple faces tracking (GMFT) algorithm in complex background. In GMFT method first faces are detected by combination of morphological facial feature extraction and gradient-based edge detection methods. After a face is reliably detected, it is tracked over time with a novel real-time algorithm. The algorithm has been implemented and tested under a wide range of real-world conditions. The resulting system runs in real-time on a standard PC, being robust to face scale variations, rotations in depth, and fast changes in subject/camera position. It has consistently provided performance which satisfies the following requirements: 1) able to automatically determine the initial position and size of face(s) and track it/them in complex background; 2) insensitive to face orientation and scale changes; 3) insensitive to pose, lighting, and expression variations.
Date of Conference: 6-8 Aug. 2007